A novel skin cancer detection architecture using tangent rat swarm optimization algorithm enabled DenseNet.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• Skin cancer poses a significant global health challenge, necessitating improved diagnostic tools.
• Early and accurate detection is crucial for reducing mortality and morbidity rates associated with skin cancer.
• Current diagnostic methods can be subjective and time-consuming, highlighting the need for automated, objective approaches.

Purpose of the Study:

• To develop and evaluate a novel deep learning model, TRSO-DenseNet, for accurate skin cancer detection.
• To enhance the performance of the DenseNet architecture through structural optimization using the Tangent Rat Swarm Optimization algorithm.
• To improve the accuracy and efficiency of skin cancer diagnosis using automated image analysis.

Main Methods:

• Pre-processing of input images using a linear smoothing filter.
• Skin lesion segmentation utilizing Mask-RCNN, followed by extensive image augmentation.
• Feature extraction incorporating statistical, Haralick texture, CNN, Local Ternary Pattern (LTP), Histogram of Oriented Gradients (HOG), and Local Vector Pattern (LVP) features.
• Optimizing DenseNet architecture with Tangent Rat Swarm Optimization (TRSO), a hybrid of Tangent Search Algorithm (TSA) and Rat Swarm Optimizer (RSO).

Main Results:

• The TRSO-DenseNet model achieved a high accuracy of 94.63% on the SIIM-ISIC Melanoma Classification dataset.
• The model demonstrated superior performance with a True Positive Rate (TPR) of 91.51% and a True Negative Rate (TNR) of 92.46%.
• The integration of TRSO significantly enhanced the structural optimization and detection capabilities of the DenseNet model.

Conclusions:

• The proposed TRSO-DenseNet model offers a highly accurate and effective solution for automated skin cancer detection.
• This AI-driven approach has the potential to significantly improve early diagnosis rates and patient outcomes.
• Further research and clinical validation are warranted to integrate this technology into routine dermatological practice.